Method and device for measuring motion speed and computer readable storage medium

CN118946904APending Publication Date: 2024-11-12YINWANG INTELLIGENT TECHNOLOGIES CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202380023034.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2023-02-06
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

In autonomous driving tasks, pure vision solutions based on cameras are difficult to accurately measure the speed of target objects, resulting in low speed accuracy and unable to meet practical application requirements.

Method used

By inputting the continuous images captured by the first object into the neural network, the image features of the same target object are identified and filtered, and the speed of the target object is directly calculated, reducing dependence on position information and improving speed measurement efficiency and accuracy.

Benefits of technology

It improves the measurement accuracy of the target object's movement speed, reduces the occupancy rate of computing resources, enhances the accuracy of the speed measurement results, and is suitable for simultaneous speed measurement of multiple target objects.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118946904A_ABST
    Figure CN118946904A_ABST
Patent Text Reader

Abstract

The invention discloses a method for measuring a motion speed, a computing device and a computer readable storage medium. Motion features and N shot images of the first object are obtained, the N shot images are continuous N frames of images obtained through shooting of the first object, and N is an integer larger than 1; the N shot images are input into a first neural network, N first images and M second images are obtained, each first image corresponds to the feature of one shot image, the M second images comprise the same target object, and M is an integer larger than 1 and smaller than or equal to N; and inputting the motion features, the N first images and the M second images into a second neural network to obtain the speed of the target object. According to the method, the position and coordinate information of the target object do not need to be speculated any more, the occupancy rate of computing resources is low, and therefore the efficiency of the speed measurement process is improved. Moreover, the speed of the target object is obtained directly according to the shot image of the first object and the accurate motion characteristics of the first object, so that the accuracy of the speed measurement result is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Method, device and computer-readable storage medium for measuring motion speed Technical Field

[0001] The embodiments of the present application relate to the field of artificial intelligence, and in particular to a method, device, and computer-readable storage medium for measuring motion speed. Background Art

[0002] Calculating the velocity of target objects, including vehicles and pedestrians, has always been a challenging issue in autonomous driving. Accurately estimating the velocity of a target object can aid in calculating its trajectory, thereby predicting its future behavior. This allows autonomous vehicles to make more informed driving decisions and plan safer and more efficient vehicle maneuvers.

[0003] In a purely vision-based camera solution, measuring the speed of a target object primarily involves first measuring its position, then determining its lateral and longitudinal velocities, as well as its angle of velocity, based on the distance change. The position of the target object can be regressed from the world coordinate system using a neural network, or calculated using traditional geometric methods.

[0004] However, when measuring the position of a target object through purely visual methods, there will be a certain deviation. Then, the target object speed calculated based on this deviated position will have lower accuracy and will be difficult to meet the requirements of practical applications.

[0005] Summary of the Invention

[0006] The present application provides a method for measuring motion speed, a computing device, and a computer-readable storage medium, for improving the accuracy of measuring the motion speed of a target object.

[0007] In a first aspect, the present application provides a method for measuring motion speed. A first object has a shooting function and can shoot scenes within the shooting visual range of the first object to obtain N captured images, where N is an integer greater than 1. The N captured images are N consecutive frames of images obtained by shooting the first object. Exemplarily, the N consecutive frames of images obtained by shooting the first object within a certain time period can be used as the N captured images of the present application, or, if the first object is shot in a certain time period to obtain a video, the N consecutive frames of images included in the video can also be used as the N captured images in the present application.

[0008] The first object of the present application can be in a stationary state, in a moving state, or switch between a stationary state and a moving state, and the N captured images are not limited to being captured when the first object is in a stationary state or in a moving state. The motion characteristics of the first object are used to represent the characteristics and changes in the motion state of the first object during the time period in which the N captured images are captured. For example, the motion characteristics of the first object include but are not limited to the position, coordinates, or speed of the first object.

[0009] N captured images are input into a first neural network to obtain N first images and M second images. Each first image corresponds to the features of a captured image, and the M second images include the same target object, where M is an integer greater than 1 and less than or equal to N. In this application, it is necessary to identify and filter the captured images of the same target object from the N captured images to obtain M second images. In other words, these M second images all capture the same target object. Therefore, through the M second images, it is possible to understand the features of the target object in different frames.

[0010] There may be multiple target objects within the visual range of the first object. The method for measuring motion speed in this application is applicable to measuring the speed of one or more target objects simultaneously. In scenarios where the speed of multiple target objects is to be measured, it is necessary to obtain a second image of each target object, and use the second image of the target object in the subsequent speed measurement process. For ease of description, this application only uses the example of measuring the speed of a single target object for introduction.

[0011] The motion features, N first images, and M second images are input into the second neural network to obtain the speed of the target object.

[0012] In this application, there is no need to infer the position and coordinate information of the target object, which reduces the utilization of computing resources and improves the efficiency of the speed measurement process. In addition, the speed of the target object is directly obtained based on the image captured by the first object and the accurate motion characteristics of the first object itself, thereby improving the accuracy of the speed measurement result.

[0013] Based on the first aspect, in an optional embodiment, the first neural network includes a first sub-network. After generating N corresponding first images based on N captured images, the N first images will be used as input to the first sub-network, and the first sub-network will recognize the N first images and identify the object to be measured in each first image, thereby obtaining N recognition images. At this time, among the N recognition images obtained, there will be M recognition images that recognize the same target object. For the M recognition images that recognize the same target object, they are the M second images corresponding to the target object in this application. In other words, the M second images are the recognition images that recognize the target object among the N recognition images.

[0014] Based on the first aspect, in an optional embodiment, the first neural network includes a second sub-network. After N captured images are input into the first neural network, the second sub-network firstly performs feature extraction on the N captured images to obtain N first images, each first image corresponding to the features of a captured image.

[0015] Based on the first aspect, in an optional implementation, before inputting motion features, N first images, and M second images into a second neural network, the image information in the second image except for the target object can be first shielded to obtain a shielded second image, thereby reducing interference from other redundant information in the second image (such as background streets, street lights, or buildings, etc.), thereby improving the efficiency of the speed measurement process and the accuracy of the speed measurement results.

[0016] Based on the first aspect, in an optional implementation, since the N captured images are N frames of images from consecutive different timestamps, in order to facilitate subsequent processing, the motion features of the first object can be first encoded into N motion feature images with different timestamps. The timestamps of the N captured images are the same as the timestamps of the N motion feature images, and the timestamp of each motion feature image corresponds to the timestamp of a captured image.

[0017] In a second aspect, the present application provides a speed measuring device. The speed measuring device includes:

[0018] an acquiring unit, configured to acquire motion characteristics of a first object and N captured images, where the N captured images are N consecutive frames of images captured by the first object, and N is an integer greater than 1;

[0019] a processing unit, configured to input N captured images into a first neural network to obtain N first images and M second images, wherein each first image corresponds to a feature of a captured image, and the M second images include the same target object, where M is an integer greater than 1 and less than or equal to N;

[0020] The processing unit is further used to input the motion features, the N first images, and the M second images into the second neural network to obtain the speed of the target object.

[0021] Based on the second aspect, in an optional embodiment, the first neural network includes a first subnetwork, wherein the first subnetwork is used to recognize N first images to obtain N recognition images, and the M second images are recognition images in which the target object is recognized among the N recognition images.

[0022] Based on the second aspect, in an optional implementation, the first neural network further includes a second sub-network, and the second sub-network is used to extract features from the N captured images to obtain N first images.

[0023] Based on the second aspect, in an optional implementation, before inputting the motion features, the N first images, and the M second images into the second neural network, the processing unit is further configured to:

[0024] Image information other than the target object in the second image is shielded to obtain a shielded second image.

[0025] Based on the second aspect, in an optional implementation, the motion features, N first images, and M second images are input into a second neural network to obtain the velocity of the target object, including:

[0026] The motion features, N first images, and M second images after masking are input into the second neural network to obtain the speed of the target object.

[0027] Based on the second aspect, in an optional embodiment, the motion characteristics of the first object are N motion feature images with different timestamps, wherein the timestamps of the N captured images are the same as the timestamps of the N motion feature images, and the timestamp of each motion feature image corresponds to the timestamp of a captured image.

[0028] In a third aspect, a computing device is provided, comprising a memory and a processor coupled to the memory; the memory is used to store instructions, and the processor is used to execute the instructions to implement the method described in any one of the above aspects.

[0029] In a fourth aspect, a computer-readable storage medium is provided, wherein a computer program is stored in the computer-readable storage medium, and when the computer program is executed on a processor, the method described in any one of the above aspects is implemented.

[0030] In a fifth aspect, a computer program product or a computer program is provided, which includes computer instructions and, when executed on a processor, implements the method described in any one of the above aspects. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without any creative work.

[0032] Figure 1 is a schematic diagram of the structure of the artificial intelligence main framework;

[0033] FIG2 is a flow chart of a method for measuring motion speed in the present application;

[0034] FIG3 is a visual evaluation result of the solution for measuring motion speed in this application;

[0035] FIG4 is another visual evaluation result of the solution for measuring motion speed in this application;

[0036] FIG5 is another visual evaluation result of the solution for measuring motion speed in this application;

[0037] FIG6 is another flow chart of the method for measuring motion speed in the present application;

[0038] FIG7 is a schematic structural diagram of a speed measuring device provided in an embodiment of the present application;

[0039] FIG8 is a schematic diagram of the structure of the computing device involved in the embodiments of the present application. DETAILED DESCRIPTION

[0040] Embodiments of the present application provide a method for measuring motion speed, a computing device, and a computer-readable storage medium, for improving the accuracy of measuring the motion speed of a target object.

[0041] The embodiments of the present application are described below in conjunction with the drawings in the embodiments of the present application. The terms used in the implementation methods of the present application are only used to explain the specific embodiments of the present application and are not intended to limit the present application. It is known to those skilled in the art that with the development of technology and the emergence of new scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0042] In this application, "at least one" means one or more, and "more" means two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, c can be single or plural.

[0043] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the numbers used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can, for example, be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0044] Some nouns or terms used in this application are explained below, and these nouns or terms are also considered part of the invention content.

[0045] First, let's describe the overall workflow of an AI system. See Figure 1, which shows a schematic diagram of the AI ​​framework. This framework will be explained from two perspectives: the "intelligent information chain" (horizontal axis) and the "IT value chain" (vertical axis). The "intelligent information chain" reflects the entire process from data acquisition to processing. For example, it could be the general process of intelligent information perception, intelligent information representation and formation, intelligent reasoning, intelligent decision-making, and intelligent execution and output. Throughout this process, data undergoes a condensed journey from "data-information-knowledge-wisdom." The "IT value chain," spanning the underlying infrastructure of human intelligence, information (provided and processed by technology), and the system's industrial ecosystem, reflects the value that AI brings to the information technology industry.

[0046] (1) Infrastructure.

[0047] The infrastructure provides computing power for AI systems, enabling communication with the outside world and supporting this through a foundational platform. Communication with the outside world occurs through sensors; computing power is provided by intelligent chips, such as central processing units (CPUs), embedded neural network processing units (NPUs), graphics processing units (GPUs), application-specific integrated circuits (ASICs), and field programmable gate arrays (FPGAs), among other hardware acceleration chips; the foundational platform includes distributed computing frameworks and networks, along with other related platform guarantees and support, including cloud storage and computing, and interconnected networks. For example, sensors communicate with the outside world to acquire data, which is then fed into the intelligent chips within the distributed computing system provided by the foundational platform for computation.

[0048] (2)Data.

[0049] Data above the infrastructure layer represents data sources for AI. This data includes graphics, images, voice, and text, as well as IoT data from traditional devices. This includes business data from existing systems and sensor data such as force, displacement, liquid level, temperature, and humidity.

[0050] (3)Data processing.

[0051] Data processing generally includes data training, machine learning, deep learning, search, reasoning, decision-making, etc.

[0052] Among them, machine learning and deep learning can symbolize and formalize data for intelligent information modeling, extraction, preprocessing, and training.

[0053] Reasoning refers to the process of simulating human intelligent reasoning in computers or intelligent systems, using formalized information to perform machine thinking and solve problems based on reasoning control strategies. Typical functions are search and matching.

[0054] Decision-making refers to the process of making decisions after intelligent information is reasoned, and usually provides functions such as classification, sorting, and prediction.

[0055] (4) General ability.

[0056] After the data has undergone the data processing mentioned above, some general capabilities can be further formed based on the results of the data processing, such as algorithms or a general system, for example, translation, text analysis, computer vision processing, speech recognition, image recognition, etc.

[0057] (5) Smart products and industry applications.

[0058] Smart products and industry applications refer to the products and applications of artificial intelligence systems in various fields. They are the encapsulation of the overall artificial intelligence solution, which productizes intelligent information decision-making and realizes practical application. Its application areas mainly include: smart terminals, smart transportation, smart medical care, autonomous driving, smart cities, etc.

[0059] It should be understood that the application scenarios of the method for measuring motion speed, computing device and computer-readable storage medium provided in this application include but are not limited to the above examples.

[0060] For example, in autonomous driving, calculating the velocity of objects, including vehicles and pedestrians, has always been a challenging problem. Accurately estimating the velocity of an object can aid in calculating its trajectory, thereby predicting its future behavior. This allows autonomous vehicles to make more informed driving decisions and plan safer and more efficient vehicle movements.

[0061] Existing solutions for measuring the speed of target objects can be categorized by sensor type, including methods based on lidar, millimeter-wave radar, and visual cameras. Of these, only millimeter-wave radar can directly determine the speed of moving targets, but this requires integration with vision to determine the target category. Lidar can relatively accurately determine the position of targets within a certain distance, but it is less accurate for critical targets at close range or at low speeds. Furthermore, lidar's disadvantages include its high cost, poor robustness in variable weather conditions, and its inability to meet the long-range detection requirements of autonomous driving scenarios.

[0062] In a purely vision-based camera solution, measuring the speed of a target object primarily involves first measuring its position, then determining its lateral and longitudinal velocities, as well as its angle of velocity, based on the distance change. The position of the target object can be regressed from the world coordinate system using a neural network, or calculated using traditional geometric methods.

[0063] However, when measuring the position of a target object through purely visual methods, there will be a certain deviation. Then, the target object speed calculated based on this deviated position will have lower accuracy and will be difficult to meet the requirements of practical applications.

[0064] In view of this, the present application provides a method for measuring motion speed, a computing device, and a computer-readable storage medium for improving the accuracy of measuring the motion speed of a target object. Specifically, the method for measuring motion speed of the present application is a pure visual solution based on a camera, which uses a first object to capture an image and identify the motion speed of the target object in the image. The method can be executed by the first object, or by a component of the first object (such as a processor, chip, or chip system, etc.), or by a logical node, logic module, or software that can realize all or part of the terminal functions. Please refer to Figure 2, which is a flow chart of the method for measuring motion speed in the present application. As shown in Figure 2, the method for measuring motion speed in the present application includes:

[0065] 101. Obtain motion characteristics of a first object and N captured images.

[0066] The first object has a photographing function and can photograph scenes within the first object's visual range to obtain N captured images, where N is an integer greater than 1. The N captured images are N consecutive frames of images captured by the first object. For example, the N consecutive frames of images captured by the first object within a certain time period can serve as the N captured images in this application. Alternatively, if the first object is captured in a certain time period to obtain a video, the N consecutive frames of images included in the video can also serve as the N captured images in this application.

[0067] The first object of the present application can be in a stationary state, in a moving state, or switch between a stationary state and a moving state, and the N captured images are not limited to being captured when the first object is in a stationary state or in a moving state. The motion characteristics of the first object are used to represent the characteristics and changes in the motion state of the first object during the time period in which the N captured images are captured. For example, the motion characteristics of the first object include but are not limited to the position, coordinates, or speed of the first object.

[0068] Since the N captured images are N consecutive frames with different timestamps, to facilitate subsequent processing, the motion characteristics of the first object can be first encoded into N motion characteristic images with different timestamps. The timestamps of the N captured images are the same as the timestamps of the N motion characteristic images, and the timestamp of each motion characteristic image corresponds to the timestamp of one captured image. The motion characteristics of the first object obtained after encoding (i.e., the N motion characteristic images) are used to execute step 103.

[0069] It should be understood that the term "first object" is merely a general term for devices that implement the method for measuring motion speed and the speed measurement function of the present application, and does not specifically refer to one or more devices. In actual applications, the device that implements the method for measuring motion speed and the speed measurement function of the present application may not be referred to as the "first object" but may be replaced by other names. This is not limited here. In the embodiments of the present application, only the "first object" is used as an example for illustration. For example, the first object can be a fixed-point speedometer, a drone with a camera function, or a car with a camera function, or it can be other forms of entities, which are not limited in this application. Specifically, when the first object is a car, it can specifically be a vehicle to everything (V2X) device, such as a smart car (or intelligent car), a digital car, an unmanned car (or driverless car or pilotless car or automobile), an automatic car (or self-driving car or autonomous car), a pure electric vehicle (pure EV or Battery EV), a hybrid electric vehicle (HEV), a range extended EV (REEV), a plug-in hybrid electric vehicle (PHEV) or a new energy vehicle (new energy vehicle), etc.

[0070] 102. Input the N captured images into a first neural network to obtain N first images and M second images.

[0071] N captured images are input into a first neural network to obtain N first images and M second images. Each first image corresponds to the features of a captured image, and the M second images include the same target object, where M is an integer greater than 1 and less than or equal to N. In this application, it is necessary to identify and filter the captured images of the same target object from the N captured images to obtain M second images. In other words, these M second images all capture the same target object. Therefore, through the M second images, it is possible to understand the features of the target object in different frames.

[0072] There may be multiple target objects within the visual range of the first object. The method for measuring motion speed in this application is applicable to measuring the speed of one or more target objects simultaneously. In scenarios where the speed of multiple target objects is to be measured, it is necessary to obtain a second image of each target object, and use the second image of the target object in the subsequent speed measurement process. For ease of description, this application only uses the example of measuring the speed of a single target object for introduction.

[0073] In one possible implementation, the first neural network includes a second sub-network. After N captured images are input into the first neural network, the second sub-network firstly extracts features from the N captured images to obtain N first images, each of which corresponds to the features of a captured image.

[0074] In one possible implementation, the first neural network includes a first sub-network. After generating N corresponding first images based on N captured images, the N first images will be used as input to the first sub-network, which will recognize the N first images and identify the object to be measured in each first image, thereby obtaining N recognition images. At this time, among the N recognition images obtained, there will be M recognition images that recognize the same target object. For the M recognition images that recognize the same target object, they are the M second images corresponding to the target object in this application. In other words, the M second images are the recognition images that recognize the target object among the N recognition images.

[0075] In practical applications, after obtaining N identification images, Hungarian matching can be performed on each of the N identification images to screen out M identification images that capture the same ID (i.e., the same target object), i.e., the M second images. Alternatively, other methods can be used to screen or filter out the M second images that capture the same target object, which is not limited in this application.

[0076] 103. Input the motion feature, the N first images, and the M second images into a second neural network to obtain the velocity of the target object.

[0077] The motion features, N first images, and M second images are input into the second neural network to obtain the speed of the target object.

[0078] In this application, there is no need to infer the position and coordinate information of the target object, which reduces the utilization of computing resources and improves the efficiency of the speed measurement process. In addition, the speed of the target object is directly obtained based on the image captured by the first object and the accurate motion characteristics of the first object itself, thereby improving the accuracy of the speed measurement result.

[0079] Furthermore, after inputting the motion features, N first images and M second images into the second neural network, the image information in the second image except for the target object can be shielded first to obtain a shielded second image, thereby reducing the interference of other redundant information in the second image (such as background streets, street lights or buildings, etc.), thereby improving the efficiency of the speed measurement process and the accuracy of the speed measurement results.

[0080] A visual evaluation of the speed measurement results of the motion speed measurement scheme in this application was performed. For details, please refer to the visual evaluation results shown in Figures 3, 4, and 5. In Figures 3, 4, and 5, the rectangular boxes on the right are the speed measurement results of the motion speed measurement scheme in this application, and the arrows on the rectangular boxes are the actual label values ​​of the target objects.

[0081] On the other hand, the speed measurement results of the motion speed measurement scheme in this application were quantitatively evaluated based on the three sigma criterion. Specifically, the quantitative evaluation is as follows (x is the lateral speed, z is the longitudinal speed, unit is m / s):

[0082] abs velocity err 1sigma:x:1.3513,z:1.6791;

[0083] abs velocity err 2sigma:x:4.2324,z:6.7870;

[0084] abs velocity err 3sigma:x:9.7436,z:16.1476.

[0085] In practical applications, N captured images and motion features of the first object are collected in real time. Generally speaking, the richer the information of the captured images and motion features, the more accurate the speed of the target object measured. Please refer to Figure 6, which is another flow chart of the method for measuring motion speed in this application. As shown in Figure 6, the first image, second image and motion features of the current frame obtained can be fused in real time. For the specific fusion process, please refer to the description of step 103 in the embodiment corresponding to Figure 2 above, which will not be repeated here. If the number of frames of the fusion result of the first image, second image and motion features currently obtained is small and fails to meet the preset frame number threshold, the fusion result of the current frame is saved first, and the first image, second image and motion features of the next frame are fused; when the number of fused feature images meets a certain number of frames, the accumulated fusion results are stacked (i.e., multi-frame feature stacking). Then, the fusion result of each frame can be identified by the above-mentioned first subnetwork, and the target object in the multi-frame fusion result is locked and correspondingly established by the Hungarian algorithm, so as to obtain the features of the same target object in the multi-frame fusion result, thereby calculating the speed of the target object.

[0086] Next, to better implement the above-mentioned solution of the embodiment of the present application, the embodiment of the present application also provides related equipment for implementing the above-mentioned solution. Specifically, please refer to Figure 7, which is a structural schematic diagram of a speed measuring device provided in the embodiment of the present application. The speed measuring device is used to execute the method executed by the first object in the embodiment shown in Figure 2. As shown in Figure 7, the speed measuring device includes:

[0087] An acquisition unit 201 is configured to acquire motion characteristics of a first object and N captured images, where the N captured images are N consecutive frames of images captured by the first object, and N is an integer greater than 1.

[0088] a processing unit 202 configured to input N captured images into a first neural network to obtain N first images and M second images, wherein each first image corresponds to a feature of a captured image, and the M second images include the same target object, where M is an integer greater than 1 and less than or equal to N;

[0089] The processing unit 202 is further configured to input the motion features, the N first images, and the M second images into a second neural network to obtain a speed of the target object.

[0090] In one possible design, the first neural network includes a first subnetwork, wherein the first subnetwork is used to recognize N first images to obtain N recognition images, and the M second images are recognition images in which the target object is recognized among the N recognition images.

[0091] In one possible design, the first neural network also includes a second sub-network, which is used to extract features from the N captured images to obtain N first images.

[0092] In one possible design, before inputting the motion features, the N first images, and the M second images into the second neural network, the processing unit 202 is further configured to:

[0093] Image information other than the target object in the second image is shielded to obtain a shielded second image.

[0094] In one possible design, the motion features, N first images, and M second images are input into a second neural network to obtain the velocity of the target object, including:

[0095] The motion features, N first images, and M second images after masking are input into the second neural network to obtain the speed of the target object.

[0096] In one possible design, the motion characteristics of the first object are N motion characteristic images with different timestamps, wherein the timestamps of the N captured images are the same as the timestamps of the N motion characteristic images, and the timestamp of each motion characteristic image corresponds to the timestamp of a captured image.

[0097] It should be noted that the information interaction, execution process, etc. between the modules / units in the speed measuring device are based on the same concept as the method embodiment corresponding to Figure 2 in the embodiment of the present application. For specific contents, please refer to the description in the method embodiment shown above in the embodiment of the present application, and will not be repeated here.

[0098] Please refer to Figure 8, which is a schematic diagram of the structure of a computing device 300 involved in an embodiment of the present application. Specifically, computing device 300 may be the speed measurement device in the embodiment shown in Figure 7 above, and is used to execute the method performed by the first object in the embodiment shown in Figure 2. Figure 8 shows a possible logical structure diagram of computing device 300, which may include, but is not limited to, at least one processor 301 and a communication port 302.

[0099] Further optionally, the apparatus may also include at least one of a memory 303 and a bus 304 . In an embodiment of the present application, the at least one processor 301 is used to control and process the actions of the computing device 300 .

[0100] In addition, the processor 301 can be a central processing unit, a general-purpose processor, a digital signal processor, an application-specific integrated circuit, a field programmable gate array or other programmable logic device, a transistor logic device, a hardware component, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of the embodiments of the present application. The processor can also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a digital signal processor and a microprocessor, and the like. Those skilled in the art will clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0101] It should be noted that the computing device 300 shown in Figure 8 can be specifically used to implement the steps implemented by the first object in the aforementioned method embodiment and to achieve the technical effects corresponding to the first object. The specific implementation methods of the computing device shown in Figure 8 can refer to the description in the aforementioned method embodiment and will not be repeated here.

[0102] The present application also provides a computer program product comprising instructions. The computer program product may be software or a program product comprising instructions that can be run on a computing device or stored on any available medium. When the computer program product is run on at least one computing device, the computer program product causes the at least one computing device to perform the method described in the embodiment shown in FIG. 2 .

[0103] The embodiment of the present application also provides a computer-readable storage medium. The computer-readable storage medium can be any available medium that can be stored by a computing device or a data storage device such as a data center that contains one or more available media. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive). The computer-readable storage medium includes instructions that instruct the computing device to execute the above-mentioned method for executing the embodiment described in Figure 2.

[0104] The neural network training device provided in the embodiment of the present application can specifically be a chip, and the chip includes: a processing unit and a communication unit. The processing unit can be, for example, a processor, and the communication unit can be, for example, an input / output interface, a pin, or a circuit. The processing unit can execute the computer-executable instructions stored in the storage unit so that the chip performs the method described in the embodiment shown in Figure 2 above. Optionally, the storage unit is a storage unit within the chip, such as a register, a cache, etc. The storage unit can also be a storage unit located outside the chip within the wireless access device, such as a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM), etc.

[0105] It should be noted that the device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed across multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. In addition, in the drawings of the device embodiments provided in this application, the connection relationship between the modules indicates that there is a communication connection between them, which can be specifically implemented as one or more communication buses or signal lines.

[0106] Through the description of the above embodiments, those skilled in the art can clearly understand that the present application can be implemented by means of software plus necessary general hardware, and of course can also be implemented by special hardware including application-specific integrated circuits, special CPUs, special memories, special components, etc. In general, all functions performed by computer programs can be easily implemented with corresponding hardware, and the specific hardware structures used to implement the same function can also be diverse, such as analog circuits, digital circuits or special circuits, etc. However, for the present application, software program implementation is a better implementation method in most cases. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a readable storage medium, such as a computer's floppy disk, USB flash drive, mobile hard disk, ROM, RAM, magnetic disk or optical disk, etc., and includes a number of instructions to enable a computer device (which can be a personal computer, training equipment, or network equipment, etc.) to execute the methods described in each embodiment of the present application.

[0107] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware, or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product.

[0108] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website, a computer, a training device or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website, a computer, a training device or a data center. The computer-readable storage medium can be any available medium that a computer can store or a data storage device such as a training device, a data center, etc. that includes one or more available media integrations. The available medium can be a magnetic medium, (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)).

Claims

1. A method for measuring the speed of movement, It is characterized in that The method comprises: Acquire motion characteristics of a first object and N captured images, where the N captured images are N consecutive frames of images captured by the first object, and N is an integer greater than 1; Inputting the N captured images into a first neural network to obtain N first images and M second images, wherein each of the first images corresponds to a feature of one of the captured images, and the M second images include the same target object, where M is an integer greater than 1 and less than or equal to N; The motion feature, the N first images and the M second images are input into a second neural network to obtain the speed of the target object.

2. The method according to claim 1, It is characterized in that The first neural network includes a first subnetwork, wherein the first subnetwork is used to recognize the N first images to obtain the N recognized images, and the M second images are recognized images of the target object among the N recognized images.

3. The method according to claim 2, It is characterized in that The first neural network also includes a second sub-network, and the second sub-network is used to extract features from the N captured images to obtain N first images.

4. The method according to any one of claims 1 to 3, It is characterized in that Before inputting the motion features, the N first images and the M second images into a second neural network, the method further includes: Image information other than the target object in the second image is shielded to obtain a shielded second image.

5. The method according to claim 4, It is characterized in that The step of inputting the motion feature, the N first images, and the M second images into a second neural network to obtain the speed of the target object includes: The motion features, the N first images and the M second images after shielding are input into a second neural network to obtain the speed of the target object.

6. The method according to any one of claims 1 to 5, It is characterized in that The motion feature of the first object is N motion feature images with different timestamps, wherein the timestamps of the N captured images are the same as the timestamps of the N motion feature images, and the timestamp of each motion feature image corresponds to the timestamp of one captured image.

7. A speed measuring device, It is characterized in that The speed measuring device comprises: an acquisition unit, configured to acquire motion characteristics of a first object and N captured images, wherein the N captured images are N consecutive frames of images captured by the first object, and N is an integer greater than 1; a processing unit, configured to input the N captured images into a first neural network to obtain N first images and M second images, wherein each of the first images corresponds to a feature of one of the captured images, the M second images include the same target object, and M is an integer greater than 1 and less than or equal to N; The processing unit is further used to input the motion feature, the N first images and the M second images into a second neural network to obtain the speed of the target object.

8. The speed measuring device according to claim 7, It is characterized in that The first neural network includes a first subnetwork, wherein the first subnetwork is used to recognize the N first images to obtain the N recognized images, and the M second images are recognized images of the target object among the N recognized images.

9. The speed measuring device according to claim 8, It is characterized in that The first neural network also includes a second sub-network, and the second sub-network is used to extract features from the N captured images to obtain N first images.

10. The speed measuring device according to any one of claims 7 to 9, It is characterized in that Before inputting the motion feature, the N first images and the M second images into the second neural network, the processing unit is further used to: Image information other than the target object in the second image is shielded to obtain a shielded second image.

11. The speed measuring device according to claim 10, It is characterized in that The step of inputting the motion feature, the N first images, and the M second images into a second neural network to obtain the speed of the target object includes: The motion features, the N first images and the M second images after shielding are input into a second neural network to obtain the speed of the target object.

12. The speed measuring device according to any one of claims 7 to 11, It is characterized in that The motion feature of the first object is N motion feature images with different timestamps, wherein the timestamps of the N captured images are the same as the timestamps of the N motion feature images, and the timestamp of each motion feature image corresponds to the timestamp of one captured image.

13. A computing device, It is characterized in that comprising a processor coupled to a memory, The memory is used to store instructions; The processor is configured to execute instructions in the memory so that the computing device performs the method according to any one of claims 1 to 6.

14. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

15. A computer program product, It is characterized in that The computer program product stores computer-readable instructions, and when the computer-readable instructions are executed by a processor, the method according to any one of claims 1 to 6 is implemented.